IP networks are to perform in a best effort manner and it is therefore essential to monitor network performance and network faults on these networks. For example, network applications may require an IP service provider (ISP) to monitor, as apart of a service level agreement (SLA) between an ISP and a user/client, performance metrics such as data packet loss, round trip time and/or inter-packet jitter (inter-packet latency in arrival time) in a network. The service provider and users/clients therefore need a way to measure network performance metrics thereby to ensure that the agreed level of service is maintained.
For example, an IP SLA is a component of the Internetworking Operating System (IOS) of a network endpoint and functions by provisioning IP SLAs probes in network endpoints, e.g., routers. The probes periodically generate active measurement traffic or probe packets from the source node (e.g., source router) to the destination nodes or network endpoints (e.g., destination routers). The destination endpoints add timestamp data to the measurement traffic and, send the measurement traffic back to the source node. Based on the received measurement traffic, an IP SLA in the source node may report on the performance metrics of the network.
Current methods that are used to measure performance metrics in networks rely on a user configuring probe packets, one by one, to each desired endpoint. This configuration may be done through the use of the Command Line Interface (CLI) or Simple Network Management Protocol (SNMP). As the configuration is manual, it is dependent on the knowledge of the user and the accuracy of the user when configuring the probe packets. Also, when using this configuration, the probe packets may not be able to deal with new network endpoints added to the network or old network endpoints that have failed or fallen away. Due to the evolution of IP networks and with the ever increasing size of such networks the manual configuration has become insufficient and inefficient.